| |
Abstract:
In American Sign Language (ASL), direct quotation is rarely
introduced by phrases such as "she said, . ." Rather, signers
indicate quotation by using referential shift, which is expressed
by facing slightly to one side, and there is a break in eyegaze
with the addressee. Under referential shift, when a first person
pronoun is used (i.e., a point to the signer's chest) the discourse
is interpreted unambiguously as a direct quote, and the spatial
locus associated with the signer's body no longer refers to the
person signing, but to the referent associated with the shift. This
experiment investigated whether deictic pronouns (first person
pronouns within a referential shift) are processed similarly to
anaphoric pronouns (non-first person pronouns directed toward
locations in signing space). The probe recognition technique was
used to investigate both antecedent activation and non-antecedent
suppression. Sentence stimuli consisted of an introductory sentence
which established two participants within the discourse, followed
by either a narrative continuation using a non-first pronoun or a
quotation continuation using a "shifted" first person pronoun (see
example below). The probe signs were either the antecedent (e.g.,
DENTIST) or the non-antecedent (e.g., ASSISTANT) and were presented
either before the pronoun (at the end of the introductory clause)
or 1000 msec after the pronoun. Twenty Deaf fluent ASL signers
participated in the experiment.
Example:
DENTISTa ASSISTANTb WORK TOGETHER ALL-MORNING,
PRONOUNa READY COFFEE BREAK.
"PRONOUN1st READY COFFEE BREAK."
English Translation:
A dentist and his assistant worked together all morning;
he (the dentist) is ready for a coffee break.
(the dentist said) "I am ready for a coffee break."
The results indicated that first and non-first person pronouns
were processed differently, despite the fact that they referred to
the same referent within the discourse. Specifically, first person
pronouns were significantly more effective at suppressing the
non-antecedent compared to non-first person pronouns (F(1,19) =
6.16, p < .05). RTs to the non-antecedent probe sign (e.g.,
ASSISTANT) were much slower after the shifted first person pronoun.
As expected with this type of baseline, antecedent activation was
maintained and did not differ significantly from the Before-Pronoun
position for either pronoun type (Gernsbacher, 1989). The finding
that shifted first person pronouns produced greater suppression was
initially surprising because we had hypothesized weaker
non-antecedent suppression. We reasoned that under referential
shift signers may conceptualize the addressee of the quoted speech
as present within signing space in the form a "surrogate"
representation (Liddell, 1995). Thus, the first person pronoun
within a referential shift might actually activate, rather than
suppress, a representation of the non-antecedent. However, the
results do not support this hypothesis. One possible explanation
for more effective suppression by shifted first person pronouns in
ASL may be that the referential cues are more explicit for these
pronouns, and Gernsbacher (1989) has found that the more explicit a
co-referential element is, the more quickly it can be used to
suppress non-antecedents. Shifted first person pronouns are
accompanied by a number of nonmanual cues (e.g., a shift in
eyegaze, head and body tilt, and a change in facial expression)
which serve to identify the referent associated with the pronoun.
Such cues are not present for the narrative continuation with the
non- first person pronoun.
Gernsbacher, M. (1989). Mechanisms that improve referential
access. Cognition, 32, 99-156.
Liddell, S. (1995). Real, surrogate, and token space: Grammatical
consequences in ASL. In K. Emmorey & J. Reilly (Eds).,
Language, Gesture, and Space, pp. 19-41, New Jersey: Lawrence
Erlbaum Associates.
|
